Among numerous presentations offered by experts whose modelling inputs have been informing Health Minister Zweli Mkhize and government’s plans on slowing the spread of the coronavirus in South Africa, actuary Nick Hudson repeated some of his concerns that the ongoing lockdown was likely to cause more harm than good.
He maintained that models predicting catastrophe and gloom appeared to be exaggerating the dangers of the virus, while underplaying the very clear and known risks associated with destroying people’s livelihoods.
Actual data from other countries, he said, suggested that the lockdown would not have much, if any, effect on the peak infection and its timing either.
Hudson rose to prominence earlier this month when he and a multi-disciplinary initiative he coordinated known as Pandemic Data and Analytics (Panda) sent an open letter to President Cyril Ramaphosa warning that the ongoing lockdown could cause 29 times more deaths than the virus itself, with each week of lockdown causing more loss of life in the long run.
Hudson, the CEO of private equity firm SANA Partners, said his group was made up of concerned professionals comprising actuaries, an economist, lawyers, a medical doctor, a data specialist and a statistics lecturer.
In their report they said: “Viruses kill. But the economy sustains lives, and poverty kills too.”
Speaking further on Thursday, Hudson started by saying he had been informed about the Covid-19 Modellers’ Symposium mere minutes before it began, and so had not been able to compile a presentation.
He then spoke off the cuff about Panda’s findings and conclusions, explaining they were working to inform government policy “not just in terms of modelling”; with their own projections they were getting to answers that were “quite different from everyone else”.
Hudson said “seven key corollaries of the standard SEIR [susceptible-exposed-infectious-recovered] models being deployed were being consistently violated” around the world, not just in South Africa.
He said the one observation that stood out was that the predictions about resurgences of infections ahead of the lifting of lockdowns in different countries had not been borne out in reality.
So instead of building yet another model, his team had tried to come up with a way to explain “the huge differences between various countries’ experiences with respect to the epidemic”.
“We think it’s a key to getting an approximate feel for what might happen in our future here in South Africa.”
A key finding of their empirical process was that they were expecting far fewer deaths in South Africa than elsewhere in the world, by partly taking into account that the country has “a much younger population”.
He also said: “We are expecting a peak in daily deaths much earlier than is being suggested by the SEIR models.”
Of the 66 countries worldwide that had so far peaked, “fully 56 peaked within a very tight range of 50 to 80 days after their first death”.
“When we look at that data, a very interesting feature emerges: the timing of that peak is completely independent of the stringency of lockdown. This is a very important finding, we think.”
Their conclusion from this was that the coronavirus R0 evolution (the rate at which the virus spreads from one person to another) was being “misinterpreted by the SEIR modellers”.
“What we’re saying is that there is no correlation between lockdown stringency and the number of days to peak.”
He said that this finding pertained “regardless of how we measure lockdown stringencies”.
“We don’t see any impact of lockdown.”
Hudson said the “missing link”, or what was being missed in SEIR analyses, was that there appeared to be a large portion of all populations worldwide that was simply not susceptible to developing severe or serious infections.
“These are people that would possibly never even test positive on a serology test [a test for the presence of antibodies] after exposure to the disease.
“In other words, what’s happening is their cellular immune systems, their T-cell systems, are doing the job of keeping the infection at bay and they never come up as cases or even infections in serology tests.”
Hudson pointed out that if there indeed was a “very large body of such people, what you’d expect is in fact what’s observed: no resurgence after the end of lockdown, serology tests that cap out at a 25% to 30% level – which is what’s been observed everywhere”.
“If that’s all correct, then models that assume much higher herd immunity thresholds, much higher attack rates, much higher ultimate prevalence rates or equilibrium rates … all of those models would misread the approach of an actual lower threshold as a reduction in R0 … owing to … whatever intervention is being assessed. We think that’s a real problem at the moment.”
He added that when he had looked at the other delegates’ presentations of models with “projections that flatten out dramatically” these were “in complete opposition to our empirical finding that lockdowns do not result in any kind of flattening or delay in peak”.
Hudson said they were “very open” to engaging with modelling teams to discuss and share their results.
“We’re not selfish about them at all. We just want the right answers to be found.”
It was important to him that South Africa should not start repeating the “mistakes” that had been made in other countries, “with massive overestimations of deaths and also of the duration of the peak”.
All the models he’d seen on Thursday were asking South Africans to behave in a way that “no other country in the world to date has behaved and to a level much more serious than the most serious countries that have been observed on an age-adjusted basis”.
“We think that’s a big problem. The stakes are enormous in this game, as our paper earlier in the month showed. The consequences of lockdown are a vast humanitarian crisis in and of themselves. The causes of that are well understood, and we have to be careful not to use models that are inconsistent with the reality that has emerged elsewhere in the world, and that cause us to stay in a lockdown situation that has terrible consequences for the population of South Africa.
“We have to remember that poverty kills, as well as viruses.”
In their paper and letter to Ramaphosa, Panda had pointed out that the economic impact of the pandemic would shorten the life expectancy of perhaps millions of South Africans.
Actuarial analysis could help provide “improved, data-led” decision-making, the group said. This would go beyond quantifying the current spread of the virus, and look at the impact of a longer-term humanitarian disaster, which would require the “next best action”.
You can watch the full modelling symposium in the video below: